fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150

Activity

1. dezren39

   added

   kind/bugSomething isn't working (prefixed form of the classic bug label)Something isn't working (prefixed form of the classic bug label)

   task/small-fixQueue packet kind: bounded fix with reproduction and targeted testsQueue packet kind: bounded fix with reproduction and targeted tests

   app/pm-stackpm-stack (Pad/Grist) in developing-today/codepm-stack (Pad/Grist) in developing-today/code

   on Oct 27, 2026
2. developing-today commented on Oct 27, 2026

   

   developing-todayBot

   on Oct 27, 2026 – with developing.today

   Contributor

   More actions

   CODE-286
3. github-actions

   assigned

   dezren39

   on Oct 27, 2026
4. dezren39

   added a parent issue on Oct 27, 2026

   • epic(dashboard): PM stack coherence - single-URL, fresh data, triage-able, hierarchical #1184
5. dezren39

   added

   status/needs-triageNot yet triaged; needs a first lookNot yet triaged; needs a first look

   priority/normalOrdinary priority (the default)Ordinary priority (the default)

   effort/sSmall: expected under 2 hours of workSmall: expected under 2 hours of work

   meta/enrichedIntake review done: acceptance criteria, links, and normalized labels addedIntake review done: acceptance criteria, links, and normalized labels added

   on Oct 28, 2026
6. dezren39 commented on Oct 28, 2026

   

   dezren39

   on Oct 28, 2026

   MemberAuthor

   More actions

   Intake review

   Fleet monitor intake pass, 2026-10-28. Verified still live this run: the issue is open, and no merged PR references it.

   Related: Found during the #1046 deploy (2026-10-26) and flagged there; no merged PR references this issue (checked this run), and the issue is still open with no comments — still live.

   Notes / risks: A deploy that reports success while serving old code undermines every later verification: checks run against the new tree on disk while users hit the old process. The version-marker half matters as much as the restart half — without it the next silent no-op is undetectable.

   Suggested approach: Compare a hash of the deployed tree (or the deployed commit) in install.sh; restart only on change; expose the build commit in /api/health if not already present and assert it post-restart with a short poll loop that fails the install on mismatch.

   Labels normalized this pass and an ## Acceptance criteria section appended to the body (original text preserved verbatim above it), formalizing the acceptance prose already in the issue. Marked meta/enriched.
7. dezren39

   added

   work/assignedQueue worker lifecycle stateQueue worker lifecycle state

   queue/claimedQueue worker lifecycle stateQueue worker lifecycle state

   status/in-progressQueue worker lifecycle stateQueue worker lifecycle state

   and removed

   status/needs-triageNot yet triaged; needs a first lookNot yet triaged; needs a first look

   on Oct 29, 2026
8. 12 remaining items

   Load more
9. dezren39

   added a commit that references this issue on Oct 29, 2026

   fix(pm-stack): serialize build version in /api/health and parse it as…

   ...

   f85e929
10. dezren39 commented on Oct 29, 2026

    

    dezren39

    on Oct 29, 2026

    MemberAuthor

    More actions

    🤖 Agent note · parked

    • Agent: queue worker (repo-worker-code), lane D
    • Session / ID: queue task 3042
    • Found via: queue claim
    • Task: PR 2432 implements fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150 (install.sh tree-hash change detection, restart only on change, PM_AGG_BUILD_VERSION marker in /api/health, loud poll-and-fail verification). Laptop tests: health + api-health-comments 18 pass; full suite 342 pass / 1 pre-existing fail (verified on clean main). Codex round on 9ee6ab2 raised 2 P1s — BOTH FIXED in f85e929 and replied on their threads: (1) /api/health handler now actually serializes view.version (+ HTTP-level test), (2) install.sh live_version() parses the pretty-printed JSON with bun instead of a fixed-string grep. New head CI basic SUCCESS, PR CLEAN.
    • Expectation: resume when Codex re-review on f85e929 lands (disposition it) and merge-queue stall ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403 clears (main still 5a34fa7): enqueue, merge, close fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150, strip labels.
    • Links: PR fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 · stall ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403
11. dezren39

    added

    queue/parkedQueue task is parked (waiting on CI, time, or an event)Queue task is parked (waiting on CI, time, or an event)

    status/parkedDeliberately paused with a recorded resume conditionDeliberately paused with a recorded resume condition

    and removed

    status/in-progressQueue worker lifecycle stateQueue worker lifecycle state

    queue/claimedQueue worker lifecycle stateQueue worker lifecycle state

    on Oct 29, 2026
12. dezren39

    added a commit that references this issue on Oct 29, 2026

    fix(pm-stack): include deployment inputs in the install no-op check (#…

    ...

    d16655c
13. dezren39 commented on Oct 29, 2026

    

    dezren39

    on Oct 29, 2026

    MemberAuthor

    More actions

    🤖 Agent note · Codex P2 fixed and pushed; parked on CI/Codex re-review and the merge-queue freeze (#2403)

    • Agent: queue worker repo-worker-code, OpenChamber scheduled lane. Task 3042 (resume of the parked task).

    • PR fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 — queue/3042-1150-install-restart-verify, head d16655c2. basic: proxy port fallbacks SUCCESS, CodeRabbit neutral, mergeStateStatus: CLEAN.

    • What this shift added (the one review finding still open on the PR, a P2): the install no-op check hashed only the copied package files, so rerunning the installer with a different TAILNET_IP, a rotated credential in the SOPS file, or an edited service unit skipped .env rendering and the unit copy while still reporting a successful verification. deploy_inputs_hash now hashes those three inputs (TAILNET_IP, the SOPS file, the unit file), stores them in $DST/.deploy-inputs-hash, and a no-op requires both the tree hash and the inputs hash to match — otherwise the run re-renders .env, re-copies the unit and restarts. The SOPS file is hashed as ciphertext, so no value is decrypted or printed during change detection, and a missing file hashes as absent so the first run that gains one counts as a change.

    • Verification: bash -n passes; a harness on the extracted functions shows the inputs hash is stable across identical runs and changes on each of a tailnet-address edit, a SOPS content change and a unit edit. Replied on the Codex thread (PRRT_kwDOJWP4786xw4PT). The TypeScript tests are unaffected (install.sh is not exercised by them).

    • Blocked from merging by ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403 (repo-wide merge-queue freeze) as well as by CI/Codex confirmation on the new head. Task 3042 is parked 30 min; on resume: confirm CI green and Codex clean on d16655c2, then re-enqueue fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432, merge, and close this issue.

    • Links: PR fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 · ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403 · task 3042
14. dezren39 commented on Oct 29, 2026

    

    dezren39

    on Oct 29, 2026

    MemberAuthor

    More actions

    🤖 Agent note · parked (Lane C shift, re-claimed 22:47Z)

    • Agent: queue worker (repo-worker-code), Lane C · Task: 3042 (small_fix, fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150 aggregator install restart/verify)
    • Verified fresh: PR fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 OPEN, MERGEABLE/CLEAN, head d16655c; required check basic: proxy port fallbacks SUCCESS, no failing/pending checks. Codex gate (§5): all findings dispositioned — P1 health-marker serialization and P1 version parsing fixed in f85e929, P2 deployment-inputs no-op hash fixed in d16655c, each replied on its thread; latest Codex review at d16655c completed with no findings.
    • Blocker: ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403 merge-queue freeze (main push restriction excludes the merge-queue service; admin config decision pending).
    • Action: parked 1h. Resume: re-enqueue PR fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 when ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403 fix lands, verify merge, close fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150.
15. dezren39 commented on Oct 29, 2026

    

    dezren39

    on Oct 29, 2026

    MemberAuthor

    More actions

    Agent note — parked (queue task 3042, PR #2432)

    PR #2432 (fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version) is ready: head d16655c2, MERGEABLE/CLEAN, required check basic: proxy port fallbacks pass, Codex review on that head has_any_findings:false / has_unresolved_findings:false (the earlier P1s were fixed in f85e9291 and replied on-thread). No code work remains for #1150.

    • Blocker: the repo-wide merge-queue ejection (ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403) — main has not moved since 5a34fa70 (10:57Z). Same gate as feat(aggregator): staleness watchdog tier items (#1148) #2431, id: rustfmt meta_client.rs + drop expect() in tags parser — unbreak main fmt/clippy (#2151) #2335 and id(encryption): surface stalled delivery — Session::stalled_delivery() (#187) #2221.
    • Not re-enqueued: every enqueue today has been dropped 2–6 minutes into AWAITING_CHECKS; re-enqueueing before the settings fix only produces noise on ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403. Task 3495 is blocked escalated-to-user on that decision.
    • Resume: when a PR lands normally → re-enqueue fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 at d16655c2, verify merge, complete 3042 and close fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150.
16. dezren39 commented on Oct 30, 2026

    

    dezren39

    on Oct 30, 2026

    MemberAuthor

    More actions

    Agent note — picked up (resume, awaiting_ci) and re-parked: CI + Codex now complete, gate unchanged

    • Queue task: 3042 (Lane A, repo-worker-code), PR fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432 at d16655c.
    • Fresh recheck: all checks COMPLETED (basic: proxy port fallbacks SUCCESS, full tier SKIPPED as designed), Codex re-review reports no unresolved findings. The PR is fully merge-ready.
    • Only blocker unchanged: merge-queue stall ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403, main frozen at 5a34fa7 (last merge 2026-10-29T10:57Z). No enqueue attempted.
    • Re-parked 4h. Resume when ci: merge queue ejects every PR — no merge since #2233 (10:57Z), entries dropped 2-6 min into AWAITING_CHECKS #2403 is fixed / main moves: re-enqueue fix(pm-stack): aggregator install.sh restarts on change and self-verifies deployed version (#1150) #2432, verify merged, complete, close fix(pm-stack): aggregator install.sh does not restart a running service - deploys silently no-op #1150.
17. dezren39

    closed this as completedin #2432on Oct 30, 2026
18. developing-today

    added

    status:done

    and removed

    status:in-review

    on Oct 30, 2026
19. dezren39

    removed

    status/parkedDeliberately paused with a recorded resume conditionDeliberately paused with a recorded resume condition

    on Nov 1, 2026